China
Africa
Explore chapters and articles related to this topic
Spatio-Temporal Multi-Application Request Scheduling in Energy-Efficient Data Centers
Published in Krzysztof Iniewski, Santosh K. Kurinec, Sumeet Walia, Energy Efficient Computing & Electronics, 2019
Haitao Yuan, Jing Bi, MengChu Zhou
The objective of admission control is to protect servers from overload and to guarantee the performance of applications. In [17], the authors provide joint response routing and request mapping in geographically distributed CDCs. The authors in [18] propose a coordinated method to provide admission control and to provision resources for multi-tier services in a shared platform. The reinforcement learning method and cascade neural networks are integrated to improve the scalability and agility of the system. In [11], the authors present a simple algorithm that drops excessive workload provided that the performance is met. However, these works only consider a single application and ignore the resource conflict among multiple applications. The revenue-based workload admission control method can judiciously admit requests by considering priority, revenue and the expected response time of different request flows.
Introduction
Published in Joseph Y.-T. Leung, Handbook of SCHEDULING, 2004
A disk is at its best when it sequentially transfers large blocks of data, so that it is operating close to its maximum transfer rate. Large block sizes are usually associated with synchronous requests, which are based on processes that generate requests to successive blocks of data. Synchronous requests commonly occur in multimedia applications, such as video-on-demand (VOD). Requests have to be completed at regular time intervals to allow glitch-free video viewing. Disk scheduling is usually round-based, i.e., a set of streams are processed periodically in a fixed order. An admission control policy may be used to ensure that processing of a new stream is possible with satisfactory performance for all streams and that this is done by taking into account buffer requirements. Discrete requests originate from an infinite number of sources and the arrival process in this case is usually assumed to be Poisson (with rate λ). In an OLTP system sources correspond to transactions executing at a high level of concurrency, which generate I/O requests during their execution. A disk access is only required when the requested data block is not cached at a level of the memory hierarchy preceding the disk.
Broadband Networks
Published in Naoaki Yamanaka, High-Performance Backbone Network Technology, 2020
The second question is particularly relevant in the transitory network growth phase when it may be difficult for a network to keep up with demand. The impact of overload depends on whether or not admission control is employed. Indeed, the absence of admission control appears incompatible with transaction pricing: overload would lead to unacceptable transparency and throughput performance degradation for a paid transaction. Use of admission control is intended to ensure that performance is acceptable for admitted transactions with overload manifested by increased blocking probabilities.
Neural adaptive admission control framework: SLA-driven action termination for real-time application service management
Published in Enterprise Information Systems, 2021
Tomasz D. Sikora, George D. Magoulas
Scheduling mechanisms for tasks and requests, resource allocation, load-balancing, auto-scaling and admission control on the cloud have been the topics of considerable research attention in recent years. Most of the research done so far has tackled the control problem using a combination of components working in an inter-dependent manner, often operating in a sequence where admission control is done first, and then scheduling and resource allocation are utilised. For example, such an approach can be found in the work of Urgaonkar and Shenoy (2004) that aims to manage CPU and network bandwidth in shared clusters. Also, Yu et al. (2008) used admission control in the context of control of virtual network embedding, while Sharifian, Motamedi, and Akbari (2008) applied admission control for a cluster of web servers together with load balancing, and Almeida et al. (2010) proposed joint admission control and resource allocation for virtualised servers. Lastly, ferrer2012optimis introduced a framework, called OPTIMIS, for cloud service provisioning where admission control is a key component.
Design of MAC Layer Resource Allocation Schemes for IEEE 802.11ax: Future Directions
Published in IETE Technical Review, 2018
Rashid Ali, Sung Won Kim, Byung-Seo Kim, Yongwan Park
The main purpose of admission control is to protect the performance of the resources by efficiently deciding on the available resources. Besides, admission control collects the available resources information, and it can be helpful for MAC-RA to decide whether there are sufficient resources available for newly arriving packets. It makes admission control one of the key elements for MAC-RA. Recently, researchers have spent plenty of their time for enhancing the techniques to utilize this element efficiently [52–56]. A proportional admission control scheme [57] to use the beacon frame for managing resources in OBSSs is already introduced by the IEEE 802.11aa amendment. The proportional admission control scheme has limitation of restricting the throughput of the OBSS to maintain the QoS.